It's 10pm.Do you know what your AI is doing?

Personas

← Personas
Bjorn persona

Bjorn

Business Owner

"I brought AI in to get ahead of the competition. So far it writes nicer emails. What keeps me up is different: if one of my people pastes a customer's file into some assistant to get the job done, where does that data go? Who is responsible? And when the customer asks, am I the one accountable? I think I am. I have nothing that shows me either way."
Small business owner. A handful of employees, a few hundred customers. Bought AI for an edge over the competition, has not seen results yet.
As Bjorn, I want to see which customer data leaves my business, through whom, to which AI service, and whether our rules were broken, so that I can answer a customer, a regulator or the insurer with a record instead of a guess, because the business is mine and "he meant well" is not a defense.
Small Business AI Guardian
Small office router, a handful of workstations, a customer system, staff phones on the Wi-Fi. Two AI subscriptions she pays for and an assistant that appeared in the last software update. Half of what touches customer data she knows about, none of it written down.
She does not know what AI is for yet, only that the competition claims to. Every customer record that leaves is hers to answer for, and the edge is worthless if it costs her a customer's trust.
A near miss, a customer asks "do you use AI with our data," the insurer's renewal form asks about AI controls, or Floki shows up with his record and a speech. Wants something that makes the worry go away, shows her what the AI tools actually do, and gives her a report to hand over.
Her people, who mean well. The customers, who will ask. The insurer, who already asks. Maybe an MSP who set up the router and could install the Guardian. No IT department.
Small Business AI Guardian

Gateway. Guard mode on workstations with customer data, watch mode on the rest, a monthly record to hand over, and a view of what the AI tools actually do.

“Which customer data left, through whom, to which AI, and whether a rule was broken. One page, monthly, yours to hand over.”
  • One box on the office network for 5 to 25 people, no agents, no admin console to learn
  • Customer data as the unit of concern: which record left, through whom, to which service
  • A monthly record written for a customer, a regulator or the insurer, not for a CISO
  • A plain view of what the paid AI tools actually do all day, so she can judge the edge
  • Installable by the owner or the MSP who set up the router
US small and mid-market surveys, 2024 to 2026

Figures reproduced from the October 26, 2026 persona one-pager and attributed there to the sources below; not independently verified.

60%
of small and mid-market owners say employees use public AI chatbots for work; 36% have a written policyNationwide / Edelman, Sept 2026
27%
have rules on what company or customer data may go into AI tools; 25% verify AI output before using it
35%
believe employees use unauthorized AI tools; 31% were targeted by a GenAI scam in the past year
82%
say they need more resources against AI-enabled attacks; 35% have a current incident plan
27%
of employees have entered confidential company data into public AI; 15% regularly enter customer dataSalesforce
11%
of data pasted into ChatGPT-type tools is sensitive; source code is the most common leakCyberhaven
57%
of US small businesses invest in AI, up from 36% in 2023; 24% among firms with 1 to 9 employeesBusiness.com
45%
of SMB workers worry that too much AI could harm the company's reputation

Takeaway: The use is there, the rules are not, and the owner knows it. She does not need a policy template. She needs to see what happens and have something to hand over.

Sources cited in the one-pager: Nationwide / Edelman Intelligent small business survey Sept 2026; Salesforce State of IT 2024 and Cyberhaven 2024 (via Airia roll-up); Business.com 2026 Small Business AI Outlook; Rhindon Cyber SMB buyer guide 2026 (vendor-authored); vendor pages.

5 adjacent options · October 2026 research notes

Offerings, prices and gaps are the one-pager’s assessment, not independently verified. Vendor claims and pricing may change.

Nudge Security
SaaS and AI app discovery from email metadata, employee nudges
$750 a month flat under 150 users
Gap: Inventory of accounts, not of data. No gateway, no record of what left
Microsoft Purview + Defender
AI usage discovery and DLP on managed devices, Edge controls
In M365 Business Premium, about $22 per user a month
Gap: Microsoft only, managed devices only, needs an admin. Invisible to a 10-person shop
Nightfall, Cyberhaven, Obsidian
Browser or endpoint DLP that blocks sensitive data reaching AI tools
$30K to 40K a year, built for 300+ seats
Gap: Priced and staffed for enterprises. Agents on every device
RAIC (Rhindon)
Browser extension plus connectors, AI registry, framework reports
From $349 a month
Gap: Compliance paperwork for MSPs. Browser only; the pasted file by phone or app is unseen
Firewalla, MSP firewalls
Small office firewall, rules, VPN, blocklists
$249 to 899 plus MSP fee
Gap: Network security, no AI view, no customer-data record